Can Alcohol Slow a Stroop Test Without Worsening Its Inhibition Score?

Yes. A person can respond more slowly after alcohol without showing a larger Stroop interference score. In a 40-person crossover experiment, alcohol increased average Counting Stroop reaction time by approximately 17 milliseconds, while the additional delay caused by conflicting information did not significantly increase. These are different outcomes, and treating either one as a complete measure of driving impairment would overstate the evidence.

Riedel and colleagues’ 2021 study examines an important distinction in DUI-related cognitive evidence: general slowing does not necessarily identify a specific failure of attentional inhibition, and a preserved inhibition score does not mean performance is unchanged.

What the Counting Stroop task measured

Participants saw groups of identical digits and reported how many digits appeared rather than what the digits meant. For example, “333” contains three digits and is congruent. “33” contains two digits even though each digit denotes three, making the trial incongruent. Four response keys corresponded to counts of one through four.

The task contained 80 trials of each type. Researchers recorded reaction times, error rates and an inverse efficiency score combining speed and accuracy. The interference score was the difference between incongruent and congruent performance. This subtraction seeks to isolate the extra cost of conflict rather than overall response speed.

Alcohol exposure was controlled with an infusion

Each participant completed alcohol and saline sessions in randomized order in a single-blind crossover design. An intravenous alcohol infusion was adjusted using computer-assisted protocols and repeated breath readings. The investigators targeted an arterial BAC approximation of 80 mg/100 mL, equivalent in numerical concentration units to 0.08 g/100 mL. The task began approximately 25 minutes after infusion started and lasted about eight minutes.

The paper uses arterial BAC terminology, but the exposure readings came from an Alcotest 6810 med breath analyzer. Its 2,100:1 blood-to-breath conversion approximated arterial alcohol during infusion. This was not a series of directly sampled arterial whole-blood ethanol measurements. Nor were these serum/plasma, urine or biomarker concentrations.

A controlled intravenous plateau differs from ordinary oral drinking with changing absorption. The experiment does not validate a personal blood-to-breath ratio or reconstruct any driver’s concentration curve.

Slower responses were a positive finding

Alcohol increased reaction time by approximately 17 milliseconds regardless of trial type; the reported main effect was significant (p = .006). The inverse efficiency score also worsened, whereas the overall error-rate effect was not significant. Separately, incongruent trials were approximately 49 milliseconds slower than congruent trials, showing that the task produced the expected interference effect.

These findings indicate a change in overall task performance. They do not identify whether the slowing arose from counting, processing speed or another mechanism. The study did not directly measure each of those possibilities.

The unchanged interference score was also meaningful

The alcohol-by-trial-type interaction for reaction time was not significant (p = .337). Corresponding interference effects for errors and inverse efficiency were also not significant. Bayesian model comparisons provided moderate support for a model without the additional alcohol-related interference effect, rather than simply leaving a nonsignificant p-value uninterpreted.

The authors estimated that their design could detect an interference effect down to approximately Cohen’s d = 0.58. Smaller effects cannot be excluded with the same confidence. Supplemental analyses did not identify alcohol-specific changes in conflict adaptation or interference across reaction-time quantiles.

One interpretation is that alcohol affected general performance while sparing the attentional inhibition captured by this particular subtraction score. An alternative, compatible explanation is that participants adopted a relatively automatic counting strategy. The findings do not prove that every form of inhibition is unaffected.

Who was studied limits generalization

The sample was selected for medium-to-high risk drinking but excluded current or previous alcohol use disorder. Thirty-six participants were men and four were women. Participants were also enrolled in a larger naltrexone-versus-placebo clinical trial; that factor was not part of the main analysis, and the paper provides additional subgroup analyses.

The task lacked neutral and switching conditions, and error rates were relatively low. The study did not directly compare the Counting Stroop task with stop-signal or Go/NoGo performance in the same experiment. It also did not include road driving or validate a roadside impairment classification.

How this helps interpret DUI evidence

A report of a “normal inhibition score” should identify exactly how the score was calculated. Similar slowing in both trial types can leave their difference unchanged. Conversely, slower total performance alone cannot establish which cognitive process failed.

This complements evidence that cognitive domains respond differently to alcohol. Here, the distinction concerns different measurements within one task. Neither a preserved interference score nor a group-average delay supplies an individual BAC, a crash-causation finding or a legal conclusion.

Primary source

Riedel P, Wolff M, Spreer M, et al. Acute alcohol does not impair attentional inhibition as measured with Stroop interference scores but impairs Stroop performance. Psychopharmacology. 2021;238:1593–1607. https://doi.org/10.1007/s00213-021-05792-0.